MG-126 Data sharing and variant classification consensus building in the canadian open genetics repository (COGR)

Abstract

BackgroundThere is a critical need for collaborative measures between Canadian institutions to better facilitate variant analysis and data sharing.ObjectivesThe Canadian Open Genetics Repository (COGR) is a collaborative effort for the collection, sharing and analysis of variants reported by medical diagnostics laboratories across Canada. The project focuses on reaching consensus agreements on variant classification among clinical laboratories through data sharing and analysis and disseminating such information to a large, public data repository.Design/methodCOGR provides laboratories a custom Variant Assessment Tool, and facilitates sharing through GeneInsight®, a database capturing variant interpretations, reference sequence data, and gene-disease associations. Agreements and discrepancies for individual variant interpretations were identified, and a voting system was put in place to attempt to reach consensus on discrepant classifications.ResultsThe COGR network currently contains over 3,000 variants across 23 genes associated with 10 diseases. There are 46 variants seen by at least three laboratories with fully consistent classifications. A total of 96 variants had discrepant classifications across at least two laboratories. When targeting the 5 most discrepant variants for review via a presentation and anonymous voting, only 1 variant reached consensus.ConclusionsBy sharing data among laboratories, consensus classifications could quickly be reached for a subset of variants. However, despite presenting evidence for variant classification and discussions amongst experts in the field, there is still considerable difficulty reaching consensus for ambiguous variants. This highlights the need for structured and rule-based variant review. The COGR is facilitating collaboration between Canadian laboratories and international efforts through data sharing and consensus building.